Inflammation is an important component of physiological host defense. Increasingly, however, it is clear that temporally or spatially inappropriate inflammatory responses play a part in a wide range of diseases, including those with an obvious leukocyte component (such as autoimmune diseases, asthma or atherosclerosis) but also in diseases that have not traditionally been considered to involve leukocytes (such as osteoporosis or Alzheimer's disease).
The chemokines are a large family of signalling molecules with homology to interleukin-8, which have been implicated in regulating leukocyte trafficking both in physiological and pathological conditions. With more than fifty ligands and twenty receptors involved in chemokine signalling, the system has the requisite information density to address leukocytes through the complex immune regulatory processes from the bone marrow, to the periphery, then back through secondary lymphoid organs. However, this complexity of the chemokine system has at first hindered pharmacological approaches to modulating inflammatory responses through chemokine receptor blockade. It has proved difficult to determine which chemokine receptor(s) should be inhibited to produce therapeutic benefit in a given inflammatory disease.
More recently, a family of agents which block signalling by a wide range of chemokines simultaneously has been described: Reckless et al., Biochem J. (1999) 340:803-811. The first such agent, a peptide termed “Peptide 3”, was found to inhibit leukocyte migration induced by 5 different chemokines, while leaving migration in response to other chemoattractants (such as fMLP or TGF-beta) unaltered. This peptide, and its analogs such as NR58-3.14.3 (i.e., c(DCys-DGln-DIle-DTrp-DLys-DGln-DLys-DPro-DAsp-DLeu-DCys)-NH2), are collectively termed “Broad Spectrum Chemokine Inhibitors” (BSCIs). Grainger et al., Biochem. Pharm. 65 (2003) 1027-1034 have subsequently shown BSCIs to have potentially useful anti-inflammatory activity in a range of animal models of diseases. Interestingly, simultaneous blockade of multiple chemokines is not apparently associated with acute or chronic toxicity, suggesting this approach may be a useful strategy for developing new anti-inflammatory medications with similar benefits to steroids but with reduced side-effects.
However, peptides and peptoid derivatives such as NR58-3.14.3, may not be optimal for use in vivo. They are quite expensive to synthesise and have relatively unfavorable pharmacokinetic and pharmacodynamic properties. For example, NR58-3.14.3 is not orally bioavailable and is cleared from blood plasma with a half-life period of less than 30 minutes after intravenous injection.
Two parallel strategies have been adopted to identify novel preparations that retain the anti-inflammatory properties of peptide 3 and NR58-3.14.3, but have improved characteristics for use as pharmaceuticals. Firstly, a series of peptide analogs have been developed, some of which have longer plasma half-lives than NR58-3.14.3 and which are considerably cheaper to synthesise. Secondly, a structure: activity analysis of the peptides has been carried out to identify pharmacophores in order to propose small non-peptidic structures which might retain the beneficial properties of the original peptide.
This second approach yielded several structurally distinct series of compounds that retained the anti-inflammatory properties of the peptides, including 16-amino and 16-aminoalkyl derivatives of the alkaloid yohimbine, as well as a range of N-substituted 3-aminoglutarimides. (Reference: Fox et al., J Med Chem 45(2002) 360-370; WO 99/12968 and WO 00/42071). All of these compounds are broad-spectrum chemokine inhibitors which retain selectivity over non-chemokine chemoattractants, and a number of them have been shown to block acute inflammation in vivo.
The most potent and selective of these compounds was (S)-3-(undec-10-enoyl)-aminoglutarimide (NR58,4), which inhibited chemokine-induced migration in vitro with an ED50 of 5 nM. However, further studies revealed that the aminoglutarimide ring was susceptible to enzymatic ring opening in serum. Consequently, for some applications (for example, where the inflammation under treatment is chronic, such as in autoimmune diseases) these compounds may not have optimal properties, and a more stable compound with similar anti-inflammatory properties may be superior.
As an approach to identifying such stable analogs, various derivatives of (S)-3-(undec-10-enoyl)-aminoglutarimide have been tested for their stability in serum. One such derivative, the 6-deoxo analog (S)-3-(undec-10-enoyl)-tetrahydropyridin-2-one, is completely stable in human serum for at least 7 days at 37° C., but has considerably reduced potency compared with the parental molecule.
One such family of stable, broad spectrum chemokine inhibitors (BSCIs) are the 3-amino caprolactams, with a seven-membered monolactam ring (see, for example, WO2005/053702 and WO2006/134385). However, further useful anti-inflammatory compounds have also been generated from other 3-aminolactams with different ring size (see for example WO2006/134385). Other modifications to the lactam ring, including introduction of heteroatoms and bicyclolactam ring systems, also yield compounds with BSCI activity (see, for example, WO2006/018609 and WO2006/085096).
To date, the identification of broad classes of agents with BSCI activity, and hence anti-inflammatory properties in vivo, has been based on optimising potency of the BSCI activity. For example, previous disclosures taught that introduction of 2,2-disubstitution (at the alpha- or key-carbon atom in the acyl side chain of acyl-3-aminolactams) leads to a considerable increase in potency as a BSCI, both in vitro and in vivo in models of acute inflammation, whether the 2,2-disubstituted acyl group was open chain (see WO2005/053702), monocyclic (see WO2006/134384) or polycyclic (see WO2006/016152).
However, potency of the desired pharmacological effect is only one factor in determining whether an agent will make a useful human pharmaceutical, albeit an important factor. In particular, the pharmacokinetics (or disposition of the agent within the body) exerts a major effect on the utility of a particular agent. Pharmacokinetics (defined in its broadest sense, as the study of the effects of the body on the drug, in contrast to pharmacodynamics, which is the study of the effects of the drug on the body) depends on a host of complex physiological processes, including (but not limited to) absorption, plasma stability, volume of distribution (and in particular rate of equilibration into ‘target’ tissues), metabolic transformation (including hepatic metabolism, such as cytochrome P450 isoenzyme mediated oxidation, and phase II metabolism such as sulfation and glucuronidation, and extrahepatic metabolism, such as serum enzymic modification), and excretion (such as renal clearance into urine and fecal elimination). These processes are often collectively referred to as the ‘ADME’ properties of the agent (ADME being an acronym for Absorption, Distribution, Metabolism and Excretion).
Another important factor in determining the utility of an agent as a human pharmaceutical is safety. Many, if not all, compounds administered elicit multiple effects on the body of which the desirable pharmacological effects are usually only a subset. The remaining effects may result in harm (toxic effects) or inconvenience (side-effects) to the patient. The study of such properties of candidate pharmaceutical agents is called toxicology or safety pharmacology. Unwanted effects can be broadly classified into two types. Class effects are intimately tied up with the desired pharmacological action, and (to a greater or less extent) are an inevitable consequence of manipulating the chosen molecular target. For example agents designed to prevent pathological inflammation will, to a degree, result in immunosuppression and an increased risk of infection. This is because inflammatory tissue damage and infection are both inextricably linked to the degree of immune system activity. As a result, all molecules sharing the identical pharmacological target will, to a greater or lesser extent, share class effects. In contrast, compound effects are specifically associated with a particular compound structure, usually as a result of an (often unexpected) interaction with a target distinct from the intended pharmacological target. In principle, it is possible to find another molecule with the same intended pharmacological effects but which is completely devoid of the compound-specific side-effects. Some compound effects are common (such as hERG interaction, which can result in dangerous prolongation of the QT interval during heart pacing, resulting potentially fatal cardiac arrythmias), while other compound effects may be apparently unique to the particular compound.
Crucially, despite decades of pharmaceutical development experience, there is still no generally accepted method for predicting either the ADME and pharmacokinetic properties of an agent, or its toxicology and safety pharmacology. It is for this reason that explicit testing, first using in vitro assay systems (such as hERG-expressing cell lines), then in animals and finally in phase I clinical trials in man, is a regulatory requirement worldwide for the development of a new pharmaceutical.
Methods have been described for predicting certain aspects of ADME from inspection of the molecular structure, and there can be little doubt that experienced medicinal chemists can reliably rule out many structures on purely theoretical grounds. An example of such a “rule of thumb” (for it is no more dependable than that) would be Lipinsky's “Rules of Five”, based on the observation that most approved pharmaceuticals meet certain criteria related to molecular weight, number of rotatable bonds and polarity. Similarly, it is generally well known that molecules with large, hydrophobic groups are more likely to show an undesirable interaction with the hERG channel.
Such general guidelines, even when applied together, may be useful for eliminating unsuitable molecules but many very unsuitable molecules (for various reasons) would still slip though the net. Today, no-one would seriously countenance selecting a drug candidate from a class of active compounds on purely theoretical grounds. As a result, the discovery of a particular compound from within a class which has particularly advantageous ADME, pharmacokinetic, toxicological and safety pharmacological properties requires considerable practical experimentation among good candidates, and is a novel finding which could not be predicted even by those skilled in the art.